0
$\begingroup$

So I am doing a biostats course on Stanford online, and I hit a snag. This is related to conditional probability would love it if anybody could explain the workings behind the process of the first image, which I assume is the basis for the second. The first image:: (sorry couldn't directly post it).

Anyways, here, the BC refers to the prevalence of breast cancer in a population. That probability tree basically means that .003 (.3%) has breast cancer (first event). The next branch is a conditional probability. The false-negative probability is .1 while the false positive rate is .89. I did calculate the two cases where the is a positive result, but couldn't understand the calculation in red. As much as I understand, it is the conditional probability of having breast cancer, given that a person is tested positive, but I cannot figure out how the workings come by (if that part in the red is from a formula, we haven't been taught that). Can anybody explain how that is achieved?

This is the second question in a quiz following the lecture.

I believe this question is based on that concept from the first picture, but I cannot understand it. It'd be great if anyone could clear that up for me. Thanks a lot :)

$\endgroup$
0
$\begingroup$

Quiz question: Let $X=1$ if exposed and $X=0$ if unexposed. Let $Y=1$ if pancreatic cancer and $Y=0$ if no cancer. Then:

\begin{align*} P(X=1|Y=1)=&\frac{P(Y=1|X=1)P(X=1)}{P(Y=1)}\\ =&\frac{P(Y=1|X=1)P(X=1)}{P(Y=1|X=1)P(X=1)+P(Y=1|X=0)P(X=0)}\\ &=\frac{0.0004\times 0.15}{0.0004\times 0.15+0.0001\times 0.85}\\ &=0.4138 \end{align*}

where the first line uses Bayes Rule and the second line uses the law of total probability.

The first question has the exact same steps: use Bayes Rule and the law of total probability, then plug-in the relevant probabilities.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.